Artificial meerkat algorithm: a new metaheuristic algorithm for solving optimization problems

Abstract In this study, a novel artificial meerkat optimization algorithm (AMA) is proposed to simulate the cooperative behaviors of meerkat populations. The AMA algorithm is designed with two sub-populations, multiple search strategies, a multi-stage elimination mechanism, and a combination of information sharing and greedy selection strategies. Drawing inspiration from the intra-population learning behavior, the algorithm introduces two search mechanisms: single-source learning and multi-source learning. Additionally, inspired by the sentinel behavior of meerkat populations, a search strategy is proposed that combines Gaussian and Lévy variations. Furthermore, inspired by the inter-population aggression behavior of meerkat populations, the AMA algorithm iteratively applies these four search strategies, retaining the most suitable strategy while eliminating others to enhance its applicability across complex optimization problems. Experimental results comparing the AMA algorithm with seven state-of-the-art algorithms on 53 test functions demonstrate that the AMA algorithm outperforms others on 71.7% of the test functions. Moreover, experiments on challenging engineering optimization problems confirm the superior performance of the AMA algorithm over alternative algorithms.